Identifying Influential Nodes And Immunization In Complex Networks Based On Dynamical Strategy

Since many real systems can be abstracted as complex networks,the study on the complex network has become a hot field.As the two important topics--identifying influential nodes and immunization,they are getting more and more attention.Traditionally,study on the two problems is based on the network topology information,however,with the popularity of mobile Internet and Internet of things,it is more convenient to obtain the dynamical unstructured information than the whole structural information.Inspired by this problem,this thesis includes the following two parts:1.Based on the label spreading dynamics,we propose a centrality index to identify influential nodes in complex networks,where the influence of a node is measured by how many different labels who have received.Under different spreading models,we compare our index with several traditional centrality indices in different networks,our results indicate that the performance of our index is better than others.2.As for the seasonal epidemics in complex network,we proposed an immunization strategy with memory effort.The main idea is that the proba-bility of immunization for each node is not only proportional to the number of infected neighbors in the current season but also proportional to the number of the infected neighbors in the past seasons.Our algorithm is validated by the SIR spreading model in different networks.